Today!
TOMORROW: Join us at 2 pm ET for @sejournal's live webinar with @THCapper on AI-Organized SERPs & Overviews.
Save your spot: bit.ly/3E6ZEnO 👈
3
570
Including an analysis of the untypical DA profile of training sets created specifically for LLMs.
New data from @ziffdavis (owners of CNET, IGN, PCMag, Mashable and more) shows that AI companies are using more premium publisher data than they care to admit.
axios.com/2024/11/05/ai-trai… by @sarafischer via @axios
I had a theory from the Google leaks: The HCU is not about content.
It's about brand:links ratio.
I've looked at the last 3 updates, ~2m keywords, and tens of thousands of sites, trying to prove it.
Results are a lot clearer than I expected.
Who is blocking AI bots, and why?
moz.com/blog/block-ai-bots xitter.twitterapi.xyz/THCapper/status/181862…
I wrote a thing about (imo) the most interesting part of the Google leak. Yes, it isn't revolutionary - I even wrote about these concepts in 2018. But understanding how these critical parts fit together is:
a) Very satisfying
b) Something we can build on
moz.com/blog/google-navboost…
I'm particularly interested in @iPullRank's linking of a patent associated with Panda to "navboost queries" (implied navigational search?).
Lines up very well with anecdotal experience of HCU and also my/@Moz's thinking with Brand Authority.
moz.com/blog/google-leak-whi… via @moz
Initial reaction Whiteboard Friday doesn't allow for a deep exploration of this idea but I (or @dr_pete) might write something up after MozCon.
190
0.34 is a massive correlation from just one factor, not adjusting for links, brand, click signals (lol), or anything else.
Remember, a score of 1 would indicate that this score perfectly predicted ranking order.
Hats off to @clearscope, well played.
"Content scores" have WEAK correlation with ranking in Google!!!!
Surprised? 😮
I wasn't. 🤷♂️
I was pretty much expecting exacly that, when I suggested to run this simple experiment to our marketing team:
1. Take a bunch of "on-page tools"
2. Take a few dozen keywords
3. Record "content scores" for each page in the SERP from each tool
4. Study correlation
You can read the full post ( by @JoshuaCHardwick & @siquanong ) here:
ahrefs.com/blog/seo-content-…
P.S.
Before you leave any comment on this tweet please DO click the link above and DO read the full article.
This post is not to say that "content scores" are useless. This post is to explain what they are and aren't good for.
For context - a comparable study of page level link scores would top out lower than this.
(For others making comparisons - needs to be 20 results per SERP to be apples to apples comparison.)
moz.com/blog/2021-links-rank…
3
195
I'm on record many times in the last year saying that SGEs/AIOs were remarkably bad, and Google couldn't possibly roll them out.
Turns out I was wrong about the last bit, but I remain baffled at how they could make a misstep this large and obvious.
Google had a year to test the AI overviews results and here we are, major media now covering their bad quality.
It seems that most of their concerns were about their monetization viability and impact rather than the quality.
Now major publications are covering it.
And no, this is not the equivalent of ranking lower quality results. It’s worse. One thing is when you rank lower quality content in some scenarios, which is already bad but at least users can easily see the source of the information.
Another is that Google directly showcases a summary with bad / wrong information packaging it in a way that makes it hard to understand where is coming from in the first place. How can users trust these answers?
Being a reliable source of information is the base of a search platform, and from what we’re seeing here, it seems that they’ve forgotten about what makes them the top search choice for the audience.
Let’s hope this makes them reconsider their approach so far.
It really does look like they're doing the worst possible version of this.
I'm baffled tbh - will wait and see how common it is and what it looks like over the next few days, but this feels like a massive unforced error.
blog.google/products/search/…
Also, in case you're wondering, yes we'll be tracking it ASAP - getstat.com/blog/sge-serp-tr…
3
158
Replacing bloggers and independent sites with household name brands may or may not be "helpful", but I could imagine Google seeing it as a safe play.
In the context of AI content / declining trust in search, few (non-SEOs) are going to argue with these kinds of result.
I think the review sites trend is an interesting comparison. In one area, Google prefers(?) big brands leveraging their brand/authority with a type of content they're not known for.
In another area, this is considered spam.
Then many SEOs think the exact opposite in both cases.
I always suggest that people aim for 5-7 minutes for a Whiteboard Friday.
Anyway, here's 13 minutes of me talking about the March 5th updates and reading between the lines: moz.com/blog/google-updates-…
Surprised I've not seen more discussion of this in SEO circles, given how popular these new robots exclusions have been.
So, I wrote up my best arguments in both directions. And, perhaps, a best (worst?) of all worlds alternative.
moz.com/blog/who-should-bloc…
What if I told you that by far the most common SERP feature in position 1 is... organic
moz.com/blog/position-one-wh…
I think this feature is also behind the YoY decline in News features we've seen, too. Which is also not a very healthy swap imo.
Serious question for you @searchliaison:
For 6+ years, and throughout the QRG, Google has shared that YMYL queries - presumably like "weight loss" - require the highest levels of E-E-A-T. For 6+ years, we've seen sites like the CDC and Harvard ranking prominently for these queries, plus Google's own "symptom search" knowledge panel, which was also supposedly curated using information from the Mayo Clinic and Harvard.
Why, then, is it ok for Google to add a "Discussions and Forums" carousel above all of these E-E-A-T-driven medical results when we are seeing very clearly how Reddit is being overtaken by affiliate spam and scammers?
This is not the only such YMYL query where Discussions and Forums ranks so prominently.
Isn't this a big dangerous for Google?
Anyone seen an SGE in the wild on a US SERP yet?
(Not on an opt-in labs account - I'm wondering what the version they're testing looks like / how it differs from labs version.)
Imo, Google contributes heavily to this.
If their messaging implies that sites losing in updates were somehow bad, measured using on-site signals (and not, e.g. user signals), it is not unreasonable for people to conclude they need to focus on "specific content things".
I'm begging. Please, I'm begging you or anyone to stop focusing on specific content things. It's not whether a TOC or saying someone is an "expert" is somehow a specific signal. It's that these things might indicate a pattern of doing too much "show Google" things that's taking away of just being a good site for people -- which is what Google wants to reward.